A Review of Tissue Substitutes for Ultrasound Imaging - McMaster

Ultrasound in Med. & Biol., Vol. 36, No. 6, pp. 861–873, 2010Copyright � 2010 World Federation for Ultrasound in Medicine & Biology

Printed in the USA. All rights reserved0301-5629/$–see front matter

asmedbio.2010.02.012
doi:10.1016/j.ultr
d Review

A REVIEW OF TISSUE SUBSTITUTES FOR ULTRASOUND IMAGING

MARTIN O. CULJAT,*yz{ DAVID GOLDENBERG,*x PRIYAMVADA TEWARI,*y and RAHUL S. SINGH*z{

*Center for Advanced Surgical and Interventional Technology (CASIT), UCLA, Los Angeles, CA; yDepartment ofBioengineering, UCLA, Los Angeles, CA; zDepartment of Surgery, UCLA, Los Angeles, CA; xDepartment of Psychobiology,

UCLA, Los Angeles, CA; and {Department of Electrical and Computer Engineering, UCSB, Santa Barbara, CA, USA

(Received 30 July 2009; revised 3 February 2010; in final form 22 February 2010)

A650 Cmculja

Abstract—The characterization and calibration of ultrasound imaging systems requires tissue-mimicking phan-toms with known acoustic properties, dimensions and internal features. Tissue phantoms are available commer-cially for a range of medical applications. However, commercial phantoms may not be suitable in ultrasoundsystem design or for evaluation of novel imaging techniques. It is often desirable to have the ability to tailor acousticproperties and phantom configurations for specific applications. A multitude of tissue-mimicking materials andphantoms are described in the literature that have been created using a variety of materials and preparation tech-niques and that have modeled a range of biological systems. This paper reviews ultrasound tissue-mimicking mate-rials and phantom fabrication techniques that have been developed over the past four decades, and describes thebenefits and disadvantages of the processes. Both soft tissue and hard tissue substitutes are explored. (E-mail:[email protected]) � 2010 World Federation for Ultrasound in Medicine & Biology.

Key Words: Tissue substitute, Tissue mimicking, Tissue equivalent, Phantom, Soft tissue, Hard tissue.

INTRODUCTION

Tissue phantoms have been used for characterization and

calibration of ultrasound imaging systems since the 1960s.

Phantoms are also used to compare the performance of

ultrasound systems for training of ultrasound technicians,

for comparison to computer models and to assist in the

development of new ultrasound transducers, systems or

diagnostic techniques. The advantage of phantoms is

that idealized tissue models can be constructed with

well-defined acoustic properties, dimensions and internal

features, thereby simplifying and standardizing the

imaging environment.

Phantoms are composed of tissue-mimicking mate-

rials, with the majority of phantoms having a simple

homogeneous internal structure. Simple or complex

targets are sometimes embedded within phantoms to

mimic internal structures or to serve as characterization

targets. Phantoms that accurately mimic heterogeneous

organs or organ systems are often referred to as anthropo-morphic phantoms. The term tissue substitute encom-

passes both phantoms and tissue-mimicking materials.

ddress correspondence to: Martin Culjat, Ph.D., CHS BH-826,harles E. Young Dr. S, Los Angeles, CA 90095. E-mail:[email protected]

861

Phantoms and anthropomorphic phantoms are avail-

able commercially, mimicking many tissues organs and

organ systems. Commercial phantoms range in price

from hundreds to thousands of dollars and are often

preferred for training and calibration of ultrasound

systems. However, commercial phantoms are typically

designed for broad markets and specific applications,

and are not customizable. For this reason, customized

design and fabrication of tissue phantoms is required for

more specialized applications requiring tailored properties

or dimensions, or when seeking to reduce cost.

This paper reviews many of the materials and tech-

niques used to prepare both soft and hard tissue-

mimicking materials and phantoms, focusing primarily

on those developed for traditional ultrasound imaging

rather than those developed specifically for elasticity

imaging (elastography), Doppler (string phantoms) or

alternate ultrasound techniques such as high-intensity

focused ultrasound (HIFU). Many of the relevant acoustic

properties and measurements are first discussed, followed

by common materials and preparation techniques used to

develop general soft tissue phantoms. The subsequent

sections focus on the development of specific soft tissue

phantoms and on the materials and techniques used to

develop hard tissues phantoms. This paper is intended to

allow the ultrasound researcher to better understand the

mailto:[email protected]

mailto:[email protected]

862 Ultrasound in Medicine and Biology Volume 36, Number 6, 2010

advantages and disadvantages of various techniques and

to select the appropriate approach for their own work.

PHANTOM AND TISSUE PROPERTIES

Tissue substitutes used in ultrasonography must

possess acoustic properties near those of the tissues of

interest, with the most critical acoustic properties of soft

tissue substitutes being the compressional speed of sound,

characteristic acoustic impedance, attenuation, backscat-

tering coefficient and nonlinearity parameter (ICRU

1998). The most relevant acoustic properties for hard

tissues include the compressional and shear wave speeds

of sound, characteristic acoustic impedance and attenua-

tion. Speed of sound, c, is typically determined by time-

of-flight measurements through a material of a given

thickness, and characteristic acoustic impedance is most

often calculated from the product of the speed of sound

and the measured density, r, of a material. The attenuation

coefficient, A, can be measured using through transmis-

sion techniques, especially for liquids and low attenuation

materials (Madsen et al. 1982). Reflective techniques may

be more appropriate for attenuation measurements at high

frequencies and for dense solids with high attenuation

(Singh et al. 2007). The backscattering coefficient, mbs,

a measure of the differential scattering cross section per

unit volume, can be measured by comparing B-mode

images between phantoms or with reference materials

(ICRU 1998). However, the backscattering coefficient is

difficult to measure accurately in most laboratories and

therefore is rarely reported in the literature (ICRU

1998). The nonlinearity parameter, B/A, is a measure of

the degree to which density in a material changes in

response to changes in pressure amplitude (Sehgal et al.

1984). However, the nonlinearity parameter of tissue

substitutes is also rarely reported. Nonlinear effects are

typically small and therefore difficult to determine using

conventional techniques (Shui et al. 2008). Also important

for tissue substitutes and phantoms is longevity, or the

period of time over which the acoustic and mechanical

properties are stable and consistent. Longevity can vary

widely, from minutes to permanence, depending on the

selected materials and preparation technique. Young’s

modulus, a measure of stiffness under isotropic condi-

tions, is an elastic property that is critical when investi-

gating phantoms for elasticity imaging. However,

Young’s modulus is not described widely throughout

the literature and therefore is not compared here. Others

have provided more thorough discussions of tissue substi-

tutes for elasticity imaging (Hall et al. 1997).

Accurate reporting of acoustic properties is highly

dependent on preparation, and handling of tissue substi-

tutes and the inherent dependence of acoustic properties,

especially attenuation and backscatter, on frequency

require that tissue substitutes mimic tissues over a broad

frequency range. Acoustic properties are usually reported

at room temperature. However speed of sound and atten-

uation in particular are highly dependent on temperature,

and the temperature dependence varies among each of the

tissue substitutes. An additional complication for the

design of phantoms is that the acoustic properties of real

tissues are not constant among people, or even within

a person’s body, and therefore the targeted acoustic prop-

erties for a given tissue are often quoted differently in the

literature. Stiffness coefficients and elastic properties of

real tissues are known to be dependent on a number of

factors, such as age, health, body location, state (in vivo,

ex vivo), fiber orientation and loading (ICRU 1998; John

2004). It is challenging, if not impossible, to create

tissue models that take all of these factors into account.

However, the variations among and within real tissues

also underscore the importance of phantoms, in that they

serve as consistent targets for calibration, ultrasound

system testing and training that cannot be provided by

human subjects, cadavers or animal models.

SOFT TISSUE-MIMICKING MATERIALS

Soft tissues are composed of muscles, tendons, liga-

ments, fascia, fat, fibrous tissue, synovial membranes,

nerves and blood vessels. Although some soft tissue phan-

toms have been developed to include many of the compo-

nents of soft tissues, the majority of tissue substitutes have

modeled each tissue as isotropic, homogeneous materials.

It is also often desirable to prepare homogeneous tissue

substitutes that mimic the broader soft tissue environment

rather than individual tissues or groups of tissues; this

approach is practical because of the relatively modest

acoustic variation among soft tissues (roughly 8% in

speed of sound when discounting marrow and tendon

[Table 1]). In addition, many tissue substitutes are made

using techniques developed for general soft tissue and

are subsequently modified to better mimic specific tissue

properties. To date, a wide range of soft tissues—in liquid,

solid or gel form—have been modeled using a variety of

formulations, described next. The acoustic properties of

these materials are summarized in Table 2. Tissue-

mimicking materials used in commercial phantoms,

including hydrogel-based Zerdine (CIRS Inc., Norfolk,

CT, USA), a condensed milk–based gel (Gammex RMI,

Middleton, WI, USA) and a urethane rubber–based

phantom (ATS Labs, St. Paul, MN, USA) are described

elsewhere (Browne et al. 2003) and are included in

Table 2 for reference.

Water and scanning gelsWater and water-based acoustic scanning gels are the

simplest tissue substitutes, with water used as a tissue

Table 1. Acoustic properties of tissues

Material Velocity (m/s) Density (kg/m3)Attenuation

(dB/cm MHz)Acoustic Impedence

(MRayl) Source

Air 330 1.2 – 0.0004 –Blood 1584 1060 0.2 1.68 ICRU 1998Bone, Cortical 3476 1975 6.9 7.38 Hoffmeister et al. 2000Bone, Trabecular 1886 1055 9.94 1.45 Wear 1999Brain 1560 1040 0.6 1.62 ICRU 1998Breast 1510 1020 0.75 1.54 ICRU 1998Cardiac 1576 1060 0.52 1.67 ICRU 1998Connective Tissue 1613 1120 1.57 1.81 Mast 2000Cornea 1586 1076 – 1.71 Mast 2000Dentin 3800 2900 80 8.0 Kossoff and Sharpe 1966Enamel 5700 2100 120 16.5 Xu et al. 2000Fat 1478 950 0.48 1.40 Mast 2000Liver 1595 1060 0.5 1.69 ICRU 1998Marrow 1435 – 0.5 – Clarke et al. 1994Muscle 1547 1050 1.09 1.62 Mast 2000Tendon 1670 1100 4.7 1.84 Hoffmeister et al. 1994Soft tissue (Average) 1561 1043 0.54 1.63 Mast 2000Water 1480 1000 0.0022 1.48 –

Tissue substitutes for ultrasound imaging d M. O. CULJAT et al. 863

substitute for medical ultrasound measurements and cali-

bration since the early days of medical ultrasound

(Robinson and Kossoff 1972). The speed of sound in

water is lower than that of soft tissue, but the addition of

7.4% ethanol by mass has been reported to increase the

speed of sound to 1540 m/s (Giacomini 1947). However,

water is limited by its low attenuation coefficient (0.2 dB/m

MHz), relative to soft tissue, and therefore does not accu-

rately mimic the soft tissue environment (Wells 1975).

Water and water-based materials are also known to

have a strong dependence on temperature, with the speed

of sound in water varying as much as 50 m/s over the

temperature range between 20–40 �C (ICRU 1998).

Despite these limitations, water will continue to be used

as a soft tissue substitute because of its ease of use and

the prevalence of immersion transducers and test tanks.

A mixture of water with glycerin or machine-cutting fluid

has also been used widely for blood and bone marrow

phantoms, and are described in more detail later.

Table 2. Acoustic properties

MaterialVelocity

(m/s)Density(kg/m3)

Attenuatio(dB/cm MH

Agarose-based 1498–16001 1016–1100 0.04–1.40

Gelatin-based 1520–1650 1050 0.12–1.5Magnesium Silicate-based 1458–1520 – 0.85Oil Gel-based 1480–1580 1040–1060 0.4–1.8Open Cell Foam-based 1540 – 0.46 dB/cm MHPolyacrylimide Gel-based 1540 1103 0.7 dB/cm @ 5Polyuerethane 1468 1130 0.13Polyvinyl Alcohol-based 1520–1610 – 0.07 – 0.35Tofu 1520 1059 0.75Water–based 1518–1574 10001 –Condensed Milk-based* 1540 – 0.5Urethane Rubber* 1460 900 0.5–0.7Zerdine* 1540 – 0.5–0.7

* Commercially available. Provided for reference.

Acoustic scanning gels, primarily composed of

water, are also used as soft tissue substitutes and typically

have higher velocities than water. However, because scan-

ning gels are designed to minimize absorption and wave

scattering, the attenuation is typically far lower than that

of soft tissue. For example, Sonotech SG Acoustic Scan-

ning Gel is specified by the manufacturer to have a speed

of sound range of 1518–1574 m/s and an impedance of

1.52–1.60 MRayl (Sonotech 2007). Attenuation data is

not available but was reported to be similar to that of

water.

Gelatin-based tissue substitutesAmong the earliest tissue-mimicking materials

prepared for ultrasound imaging were gelatin-based mate-

rials. Gelatin, a homogeneous colloid gel, is primarily

derived from collagen in animal tissues. The Madsen

group mixed gelatin with varying concentrations of

alcohol and uniformly distributed graphite powder, with

of soft tissue substitutes

nz)

Impedence(MRayl) Source

1.52–1.761 Burlew et al. 1980; Madsen et al. 1998;D’souza et al. 2000; Ramnarine et al. 2000

1.60–1.73 Madsen 1978; Bush and Hill 1983– Sheppard and Duck 19821.54–1.67 Kondo, Kitatuji 2005

z – Ophir 1981, Ophir 1984MHz 1.7 Zell et al. 2007

1.66 Kondo, Kitatuji 20051.60–1.77 Kharine, Manohar 20031.61 Wojcik, Szabo 19991.48–1.60 Giacomini 1947; Sonotech 2006– Browne et al. 20031.31 Browne et al. 2003– Browne et al. 2003


p-methyl and p-propyl benzoic acid used as preservatives

against bacterial invasion (Madsen et al. 1978; Burlew

et al. 1980). Depending on the concentration of

n-propanol in water, a speed of sound between 1520 and

1650 m/s at room temperature could be achieved. By

varying the concentration of graphite powder in the

gelatin-based compound, therefore adjusting the scat-

tering coefficient, the attenuation coefficient was varied

between 0.2 and 1.5 dB/cm at 1 MHz. These materials

were reported to have high stability near-room tempera-

ture over a period of four months, provided that the

samples were stored in a closed container below a layer

of distilled water. Reported disadvantages to this tech-

nique were instability with temperature variations, suscep-

tibility to microbial invasion and the difficulty in

achieving uniform distribution of graphite scatterers as

the particles settled during cooling (Ophir 1981).

Another gelatin-based soft tissue substitute was later

developed using gelatin and alginate and was reported to

have improved stability (Bush and Hill 1983). Bath disin-

fectants were used to minimize microbial contaminants,

and addition of calcium chloride (CaCl2) improved

thermal stability up to 25 �C. The material had a speed

of sound of 1520 m/s, with the attenuation coefficient

varying between 0.12–0.5 dB/cm MHz, with the addition

of polyethylene or lipid microspheres. Another related

study reported that a dense gelatin-alginate composition

could be embedded within the material to provide

a distinct inner structure (Bamber and Bush 1996).

Although the gelatin-alginate technique reportedly ad-

dressed the concerns related to the stability and scatterer

uniformity, this technique has not been widely adopted

in the literature.

Agarose-based tissue substitutesAgarose gel–based tissue mimics provide another

alternative to the use of graphite powders to achieve suffi-

cient attenuation and scattering properties, as well as

improved temperature resistance and particle suspension

(Madsen et al. 1998). Agarose-based techniques are the

most widely used of the soft tissue substitute preparation

techniques described in the literature. The broad use of

agarose-based substitutes is a result of their well-

characterized performance, the ease of fabrication (the

mixture can be heated in a microwave) and the flexibility

that the process provides, allowing the incorporation of

additional ingredients to achieve a range of acoustic

properties.

Agarose is derived from agar, a hydrophilic colloid

that is extracted by boiling algae. Typically, water and

propanol are mixed at a ratio designed for a targeted speed

of sound and heated. Dry, high-purity agarose is then dis-

solved into the mixture to provide structural rigidity, or

improved resistance to change in shape, while improving

thermal stability. Evaporated milk is used to increase

attenuation and is heated separately and combined with

a preservative such as thimerosal and poured into the

agarose mixture. The resulting compound congeals to

a solid mixture that can be poured into a mold. It is impor-

tant to note, however, that molds are limited to small

volume-to-surface area ratios because the congealing

substance forms a solid layer between the liquid product

and the air and prevents the remainder of the liquid from

congealing in the same manner.

The Madsen group originally developed a recipe that

resulted in a speed of sound that ranged between 1498 and

1600 m/s, density between 1016 and 1100 kg/m2 and an

attenuation between 0.04 and 1.40 dB/cm MHz (Burlew

et al. 1980). This process was later altered to include evap-

orated milk, achieving a velocity of 1540 m/s, density of

1030 kg/m2 and attenuation of 0.1–0.7 dB/cm MHz

(Madsen et al. 1998). Many agarose-based tissue substi-

tutes have subsequently been made using this technique,

with one study including glass beads to further improve

scattering properties (Burlew et al. 1980; Madsen et al.

1998, 2003; D’Souza et al. 2001). Material properties

have been reported to remain stable for as long as two

and a half years under optimal storage conditions

(Madsen et al. 1998). However, in routine laboratory

use without careful handling, longevity is often limited

to less than one month because of microbial invasion or

damage to the delicate structure.

An alternate agar-based technique, incorporated into

vascular phantoms, was recently developed as part of

a European Commission project (Ramnarine et al.

2001). Water, glycerol, benzalkonium chloride, SiC

power and Al2O3 powder were mixed with a high-

strength agar. Benzalkonium chloride was used to control

microbial invasion, Al2O3 powder to control attenuation

and SiC to vary the backscatter. The speed of sound was

reported to be 1541 m/s, the attenuation was 0.5 dB/cm

MHz and the density was 1054 kg/m3. The high-

strength agar was reported to provide superior structural

rigidity compared with standard agarose-based materials

and was well suited for vascular flow phantoms

(Ramnarine et al. 2001). A recent multi-institution study

found that the speed of sound of these materials increased

with temperatures between 22 and 37 �C at a rate of 2.1 m/

s/�C, and the attenuation decreased at a rate of 0.005 dB/

MHz/�C (Brewin et al. 2008). Frequency dependence and

longevity were also explored.

Magnesium silicate–based tissue substitutesMagnesium silicate is an inorganic substance with

a structural form that varies with applied stress. Soft

tissue-mimicking materials were created by mixing

magnesium silicate with tetrasodium pyrophosphate

(an electrolyte needed for the hardening of the gel),


n-propanol (to control the speed of sound), water and

either graphite or talcum powder (as scattering agents to

vary attenuation) (Sheppard and Duck 1982). The speed

of sound was reported to be 1458 m/s and was increased

to 1520 m/s with the addition of n-propanol. Attenuation

was measured to be 0.85 dB/cm MHz, with a linear depen-

dence on frequency reported when graphite powder was

used. Magnesium silicate–based tissue substitutes have

the advantage of temperature stability (stable from 0 to

100 �C), resistance to microbial invasion and the ability

to reform after needle biopsy procedures (Sheppard and

Duck 1982). However, these materials are not self-

supportive and therefore cannot be sculpted or molded

into predefined shapes.

Oil gel–based tissue substitutesOil gel–based tissue substitutes were developed

more recently and feature a mixture of propylene glycol,

a gelatinizer (Dibenzylidene D Sorbitol) and 10 mm poly-

methyl methacrylate (PMMA) microspheres (Kondo et al.

2005). The advantage of this technique is that the speed of

sound and attenuation increase linearly with the propor-

tion of propylene glycol. Attenuation is further varied

by increasing the impregnation of PMMA microspheres.

In addition, oil gel–based materials have the advantage

of immunity to bacterial infection. The Kondo group has

reported speeds of sound of 1480 and 1580 m/s, attenua-

tions of 0.4 and 1.8 dB/cm MHz, and densities of 1040

and 1060 kg/m3 for the nonimpregnated and impregnated

gels, respectively.

Ethylene glycol–based materials have also been

explored as tissue substitutes and have found use as

acoustic reference materials because of their uniformity

and constancy (Dong et al. 1999). However, these mate-

rials are not ideally suited as soft tissue substitutes because

of their high speed of sound (1659 m/s), density (1110 kg/m3)

and low attenuation (0.078 dB/cm at 2.25 MHz; 0.34

dB/cm at 4.5MHz).

Open cell foam–based tissue substitutesA soft tissue substitute was developed based on open

cell foam (Ophir 1981, 1984; Lerski et al. 1982), which is

composed of polyurethane foam and a salt (NaCl) water

solution. Variation of acoustic properties was achieved

using different foam materials and liquids. Altering the

concentration of NaCl changed the speed of sound,

therefore allowing the speed to be tailored to the desired

range. Ophir reported a speed of sound of 1540 m/s and

an attenuation of 0.46 dB/cm MHz (Ophir 1981, 1984).

An advantage to this process is that localized zones

mimicking tissue pathologies or variations can be

created within the material by removing regions of foam

before preparation, therefore allowing for the creation of

simple inhomogeneous phantoms. However, the

attenuation is affected strongly by changes in

temperature (4% reduction per �C) and the presence of

bubbles in the phantom. Methods to control temperature

and allow for proper wetting of the foam were described

by Ophir (1981).

Polyacrylamide gel–based tissue substitutesPolyacrylamide gels are matrix materials commonly

used in electrophoresis and are formed by the polymeriza-

tion of the acrylamide monomer. A 10% polyacrylamide

gel was used by Zell et al. (2007) as a soft tissue substitute,

and was created by mixing acrylamide:bisacrylamide with

water, a Tris/HCl buffer (pH 8.8), tetramethyethylendi-

amine (TEMED) and ammonium peroxodisulfate. The

mixture was stirred and allowed to polymerize at room

temperature for 45 minutes. Although the speed of sound

(1580 m/s) and impedance (1.7 MRayl) were within an

acceptable range for soft tissue, the attenuation was found

to be too low (0.7 dB/cm at 5 MHz). Polyacrylamide is

also highly toxic and requires special precautions during

its preparation (Zell et al. 2007).

Polyurethane tissue substitutesPolyurethane tissue substitutes have been reported to

have low Young’s modulus, good elastic recovery and

immunity from bacterial invasion (Kondo et al. 2005).

A polyurethane gel phantom was produced with a density

of 1130 kg/m3, attenuation of 0.13 dB/cm MHz and speed

of sound of 1468 m/s. The acoustic properties of polyure-

thane were shown to be dependent on molecular structure

and weight, and attenuation was shown to increase line-

arly with both temperature and frequency. However, the

molecular design of polyurethane gels is complex, and

therefore the standardization of the technique is chal-

lenging.

Polyvinyl alcohol–based tissue substitutesPolyvinyl alcohol (PVA), a synthetic polymer, has

recently been adopted as a soft tissue substitute. PVA-

based tissue substitutes have the advantage that they

have high structural rigidity, indefinite longevity, low

cost and they require fewer ingredients compared with

the more common agarose-based tissue substitutes

(Fromageau et al. 2003; Kharine et al. 2003; Surry et al.

2004).

Preparation of PVA-based tissue substitutes requires

freeze-thaw cycles to enhance cross-linking between

polymer chains. In one process, a 10%-by-weight solution

of PVA in water was frozen and thawed in 12-h cycles to

attain the desired acoustic properties, with a speed of

sound ranging between 1520 and 1560 m/s, attenuation

between 0.07 and 0.28 dB/cm MHz and impedance

between 1.60 and 1.70 MRayl (Kharine et al. 2003).

A 0.01% solution of sodium azide was used to prevent


microbial invasion. A second process was described, in

which dimethyl sulfide (DMSO), a polar aprotic solvent,

was added to facilitate structural arrangement in the solu-

tion by lowering the freezing point, thereby strengthening

the gel (Kharine et al. 2003). Properties achieved using

this technique ranged between 1550 and 1610 m/s, 0.34

and 0.35 dB/cm MHz and 1.65 and 1.77 MRayl. Glass

can be added to PVA before freezing to increase scattering

and to vary the attenuation. The primary disadvantage of

PVA-based tissue substitutes is the preparation time,

which requires multiple 12-h freeze-thaw cycles and the

requirement for precise temperature control.

Silicone polymer–based tissue substitutesSilicone products have been suggested as potential

tissue substitutes because of their longevity, stability, vari-

able Young’s modulus and their capacity to be embedded

by scatterers such as glass and plastic microspheres.

However, these materials are limited by high attenuation

and low speed of sound, reported to be ,1000 m/s

(Robertson et al. 1992; ICRU 1998).

OrganicsOrganic materials have been used as tissue substi-

tutes, including tofu and animal tissues. Tofu has an

appropriate speed of sound (1520 m/s), attenuation (0.75

dB/cm MHz) and density (1059 kg/m3) (Wojcik et al.

1999). Although tofu is low cost and does not require

preparation, it has a lower nonlinearity parameter (6.0)

than soft tissue (8.0), is susceptible to microbial invasion,

its properties cannot be adjusted and its properties vary de-

pending on the brand and preparation. A recent study also

found that the attenuation of tofu is highly frequency-

dependent, therefore limiting its use at high frequencies

(Kim et al. 2009). Porcine tissue, bovine tissue, turkey

breast and other animal tissues have been used to mimic

human tissue, but they have limited longevity and their

acoustic properties cannot be tailored (Davies and Kew

2001; Xu et al. 2005).

SOFT TISSUE PHANTOMS

Many of the soft tissue substitute preparation tech-

niques described before have been modified and tailored

to mimic specific tissues or organs. Some have combined

multiple techniques to develop phantoms or more realistic

anthropomorphic phantoms. A sampling of the techniques

used to mimic specific soft tissues and organs are

provided.

Blood phantomsBlood phantoms are used to mimic blood both acous-

tically and rheologically. Human blood itself has been

used in previous studies (Erskine and Ritchie 1985;

Weskott 1997) but is limited in its use because of its

short lifespan, damage to erythrocytes and change in

acoustic properties at room temperature (Oates 1991).

Machine cutting fluid (Syn Cut HD; Acra Tech,

Toronto, Ontario, Canada) with distilled water has been re-

ported to be a good blood-mimicking fluid (Frayne et al.

1993; Rickey et al. 1995), with viscosity similar to that

of whole blood and cellulose particles added to mimic

the backscatter of human blood (Frayne et al. 1993;

Rickey et al. 1995). This material was reported to have

a speed of sound of 1550 m/s, an attenuation of 0.2 dB/

cm MHz and an acoustic impedance of 1.6 MRayl

(Frayne et al. 1993; Rickey et al. 1995).

Many additional blood phantoms reported in the

literature have been composed of a mixture of glycerol

or glycerin and water. Sephadex, a cross-linked dextran

gel, mixed with glycerol and water, has been suggested

for use as a blood substitute (McDicken 1986; Hoskins

et al. 1990; Eriksson et al. 1991) and has been shown to

closely mimic blood if its flow is laminar (Mo and

Cobbold 1986). Another group developed a blood-

mimicking fluid using only glycerin and water (Boote

and Zagzebski 1988). The recipe was later modified by

adding polystyrene beads to improve the backscatter coef-

ficient, resulting in a speed of sound of 1600 m/s and

density of 1040 kg/m3 (Moehring and Ritcey 1996). Vari-

ations of this technique were developed using orgasol

(nylon powder; Colombes, France) particles and a surfac-

tant to create backscatter (Oates 1991; Ramnarine et al.

1999; Samavat and Evans 2006; Tortoli et al. 2006),

yeast to increase scattering (Ferrara et al. 1996) and cellu-

lose fibers for a backscatter signal comparable to human

blood (Petrick et al. 1997).

Bone marrowStudies that developed bone substitutes have also

included bone marrow. Bone marrow substitutes were

prepared using materials such as water (Moilanen et al.

2007), butter (Moilanen et al. 2004), a mixture of gelatin

and water (Clarke et al. 1994) and vegetable oil (Strelitzki

et al. 1996; Strelitzki and Truscott 1998).

Brain phantomsBecause of the comparable acoustic properties

between average soft tissue and brain tissue, various soft

tissue substitute preparation techniques can be used to

create brain phantoms. One group used an agarose-

based technique originally developed for prostate phan-

toms to create a brain phantom that was used for the

evaluation of a noninvasive focal brain surgery

(Hynynen et al. 2004). Another group used a PVA fabrica-

tion process, initially developed for soft tissue, to make an

anthropomorphic brain phantom, using a 3-D magnetic

resonance image (MRI) to create the phantom mold


(Surry et al. 2004). This technique was later adapted to

create a multilayered anthropomorphic brain phantom by

combining three different PVA layers and plastic tubing

(Reinertsen and Collins 2006).

Breast phantomsBreast phantoms that closely mimic the attenuation,

speed of sound, density and backscatter of breast tissue

were produced by the Madsen group in 1982 (Madsen

et al. 1982). The process combined oil, gelatin and agar

to mimic four layers of tissue, including glandular tissue,

adipose tissue, skin and Cooper’s ligaments (Fig. 1). The

difference in Young’s modulus between oil and gelatin

was shown to mimic the difference between healthy and

abnormal breast tissue in vivo. A different technique was

used to develop breast elastography phantoms and

featured safflower oil dispersed in a solid aqueous gelatin

(Madsen et al. 2006b). A good review of breast phantom

fabrication techniques was recently published by Madsen

et al. (2006a).

Cardiac tissue phantomsCardiac ultrasound phantoms have application to

ultrasound imaging and color-flow Doppler imaging.

One group developed a heart-mimicking phantom from

polyurethane or gelatin, portions of a commercial soft

tissue phantom and water with embedded 5-mm silica

microsphere scatterers mimicking blood (Smith and

Rinaldi 1989). This model was later improved by adding

a bifurcating aorta and three coronary arteries (Smith et al.

1991).

Eye phantomsIn an effort to determine the biomechanical proper-

ties of cornea with high-frequency ultrasound, contact

lenses were suspended in 2% agarose (Liu and Roberts

2004). The group described preliminary data in which

they simulated acoustic reflections from tissue layers

and carried out experimental studies to match the simula-

tions. However, limited details are available.

Fig. 1. Breast phantom with oil, gelatin and agar mimickingvarious layers within the phantom (Madsen et al. 1982).

Liver phantomsLiver tissue substitutes have been created using

several techniques described for general soft tissue.

Homogeneous liver phantoms have been made from

gelatin and graphite, similar to the Madsen group’s 1978

technique, to compare attenuation and echogenicity

between healthy and diseased liver (Garra et al. 1987).

Condoms filled with water have been used to mimic

cancerous and healthy livers in a 3-D ultrasound study

(Xu et al. 2003). Liver phantoms created from open cell

foam and water have been used in a study comparing

echogenic and standard biopsy needles (Hopkins and

Bradley 2001). Breast phantoms have also been used to

mimic livers during training for ultrasound-guided liver

biopsies (Nicotra et al. 1994).

Prostate phantomsThe agarose-based tissue substitute fabrication tech-

nique developed by Madsen et al. (1998) was altered to

mimic prostate tissue (D’Souza et al. 2001). Agarose-

based prostate phantoms were developed that included

water, agarose, lipid molecules, proteins, thimerosal and

glass beads. The concentrations of agarose and glass

beads were increased to increase the attenuation of the

material. Several other materials were further added to

accommodate MRI, including ethylenediamine tetraacetic

acid and Cu21 to control the longitudinal and transverse

(T1 and T2, respectively) relaxation times.

Sinus cavity phantomsDoppler ultrasound has been proposed as a technique

to diagnose sinusitis, because the viscosity of sinus fluid is

a known indicator of the presence of an infection (Jansson

et al. 2005). One group used agar with graphite powder to

construct an anthropomorphic sinus phantom using

a mold created from a human cranium. Water-glycerol

solutions with varying viscosities were used to mimic

mucous and serous fluids in the sinus (Jansson et al.

2005). A more recent study by the same group used

bovine cortical bone to cover the graphite and agar

phantom, and used milk as the fluid mimic (Jonsson

et al. 2008). Milk was selected because of the presence

of natural scattering particles.

Skeletal muscle phantomsSkeletal muscle has acoustic properties that are close

to those of average soft tissue (Table 1); therefore, many

of the general soft tissue substitute formulations can be

used to generate materials that mimic skeletal muscle.

One group briefly described a skeletal muscle phantom

that featured a gelatin-based material (Edmonds et al.

1985) and another described an agarose-based technique

for use as a multimodal phantom (D’Souza et al. 2001).

Fig. 2. Low melting point metal alloy core and container used toshape vascular flow phantoms of agar and konjac/carrageenangel (Meagher et al. 2007). The gel was poured into the containerand the metal alloy was melted away by placing the phantom in

a hot water and potassium chloride bath.


Both techniques treated skeletal muscle as an isotropic

material.

Vascular phantomsA number of techniques have been developed to

acoustically model vascular structures, with the majority

focusing on large arteries, such as the carotid artery and

coronary artery. Vascular phantoms can be grouped into

three general categories: basic vascular phantoms with

a simple tubular structure; walled vascular phantoms,

which have a closer resemblance to the arteries; and

wall-less phantoms, which do not have tubing separating

the tissue-mimicking and blood-mimicking materials.

Real vessels harvested from cadavers have also been

used as phantoms in many studies (Kerber and Heilman

1992; Dabrowski et al. 1997, 2001). However, because

of their limited longevity and variable geometries and

flow patterns, excised vascular tissues are poor models.

Basic vascular phantoms have been made from PVA

(Nadkarni et al. 2003; Schaar et al. 2005). A combination

of 10% PVA solution with 0.75% enamel paint followed

by two freeze-thaw cycles was found to have properties

similar to human vascular tissue (Nadkarni et al. 2003).

This technique allowed the elasticity to be varied by

changing the concentrations of PVA to model both

healthy and diseased tissue. Another basic vessel phantom

was constructed using latex rubber tubing to mimic the

femoral artery, and by mounting the tubing within a gelatin

filled frame to mimic the adjacent soft tissue (Zhang and

Greenleaf 2006). Another group used a rubber ring with

wires attached to the outer surface to provide fiducial

markers (Kawase et al. 2007).

Walled vascular phantoms have been built to better

understand the onset of vascular diseases using ultra-

sound. A rigid model of carotid artery bifurcation was

created by injecting water-soluble jeweler’s wax into an

acrylic mold (Bharadvaj et al. 1982). Another group adap-

ted this technique by using lead-cored nylon as fiducial

markers, acrylic and high-density polyethylene as a protec-

tive housing and layers of agar-based materials to improve

visibility of the fiducial markers in the ultrasound image

(Frayne et al. 1993). The blood-mimicking fluid was

created from machine tool–cutting fluid, as discussed

earlier. A later version replaced the agar gels with solid

polyester to improve durability but caused beam distor-

tions and artifacts because of an increased impedance

mismatch (Smith et al. 1994). In another study examining

vascular plaques in the carotid arteries, arterial phantoms

were made using an acrylic rod within a box that was filled

with a solidified mixture of agar, glycerol, distilled water

and sigma cells (Anthony and Aaron 2002; Landry and

Fenster 2002). Plaques were created by pouring the

same mixture with a reduced concentration of sigma

cells into stainless steel molds and embedding the

plaques into the acrylic rod. A water and glycerol

mixture was used as the blood substitute.

Wall-less vascular phantoms have been used for

evaluation of Doppler ultrasound systems. These phan-

toms are better suited to Doppler flow studies, because

image distortion that typically results from tube walls is

reduced (Patterson and Foster 1983; Rickey et al. 1995).

Homogeneous vascular phantoms were constructed

using the European Commission agar-based technique,

which also featured water, glycerol, benzalkonium chlo-

ride, Al2O3 and SiC (Teirlinck et al. 1998; Ramnarine

et al. 1999; Tortoli et al. 2006). A mixture of pure

water, glycerol, orgasol particles and a surfactant served

as the blood mimic (Ramnarine et al. 1999; Tortoli et al.

2006). This phantom was reported to have good

longevity and durability to flow (Ramnarine et al. 1999),

but the agar-based tissue-mimicking material was subject

to splitting at the bifurcation apex (Meagher et al. 2007).

To combat this problem, one group replaced the agar

material with konjac and carrageenan gels (Fig. 2)

(Meagher et al. 2007).

Multi-organ phantomsAnthropomorphic phantoms have been developed

that mimic complete organ systems rather than individual

tissues or organs. Madsen et al. (1980) developed a torso

section using water-alcohol–based gelatin with n-propa-

nol and various test objects to mimic the kidneys, liver,

tumors, cysts and bones. Another group designed

a multi-organ phantom for needle guidance training by


modeling various organs using balloons filled with de-

gassed water, castor oil and castor clay (Robbins 1985).

An adult female pelvis phantom was made using a hydro-

philic polymer, water, polyester fiberfill, latex and nylon

tubing (Boyce 1993). Rowan and Pederson created

a multi-organ phantom using latex to mimic skin, agar

and graphite to mimic organs and leaking silicon tubes

to simulate internal bleeding as a training tool for the diag-

nosis of internal trauma (Rowan and Pedersen 2006).

HARD TISSUE-MIMICKING MATERIALS ANDPHANTOMS

Hard tissues are mineralized tissues with a firm inter-

cellular substance and include cortical bone, trabecular

bone, dental enamel and dentin. Bone substitutes and

phantoms have been developed primarily to evaluate

and calibrate ultrasound systems designed specifically

for detecting bone pathologies (Young et al. 1993;

Clarke et al. 1994). Ultrasound imaging of teeth has not

yet become clinically accepted, but has been the subject

of various studies because of its ability to penetrate hard

tissues and its potential as a complement to radiography

(Ghorayeb et al. 2008). Dental phantoms have been

used to guide the development of dental ultrasound

imaging systems (Blodgett 2003; Culjat et al. 2005;

Singh et al. 2007). However, because of the higher

variation in acoustic properties among hard tissues, it is

more challenging to precisely match the acoustic

properties using bone phantoms and dental phantoms

than with soft tissues. On the other hand, many hard

tissue substitutes have greater structural rigidity and

longevity than soft tissue phantoms, and therefore are

more practical for long-term use. The acoustic properties

of hard tissue substitutes are provided in Table 3.

Cortical boneCortical bone, or compact bone, has a relatively

homogeneous, compact and well-defined structure.

Cortical bone substitutes have been made using epoxy,

polymers and polymer composites, with acoustic

Table 3. Acoustic properties o

Material Tissue Velocity (m/s)At

(

Acrylic Cortical Bone 2500Carbon Fiber Plastics Cortical Bone 4400Ebonite Cortical Bone 2200Epoxy Cortical Bone 2740–3168 3.7–3Perspex Cortical Bone 2657 5.3 @Epoxy Trabecular Bone 1844–3118 7–17Polyvinyl Chloride Whole Bone 2300Dental Composite Dentin 3306 108 @Aluminum Enamel 6300Soda Lime Glass Enamel 5789 6 @ 1

properties falling within the wide range of reported values

for cortical bone (Table 1). Liquid epoxy resins and hard-

eners have been mixed to create cortical bone materials,

with one group reporting a speed of sound of 3168 m/s

and attenuation of 3.7 dB/cm at 1 MHz, and another re-

porting 2740 m/s and 3.8 dB/cm at 1 MHz (Clarke et al.

1994; Tatarinov 1998). Pores were modeled for

ultrasound porosity studies by introducing 0.8–1.5–mm–

wide cubic particles of rubber in epoxy (Clarke et al.

1994; Hodgskinson et al. 1996; Tatarinov et al. 2005)

(Fig. 3). The mineral content in bone was modeled by

burning and grinding natural bone and subsequently mix-

ing the mineral residue powder into epoxy (Tatarinov

1998).

One group exploring the use of polymers and poly-

mer composites for use as cortical bone substitutes studied

various materials within the desired speed of sound range,

including ebonite (2200 m/s), acrylic (2500 m/s), carbon

fiber plastics (4400 m/s) and fiberglass (no value reported)

(Fig. 3) (Clarke et al. 1994; Hodgskinson et al. 1996;

Tatarinov et al. 2005). Perspex, a type of acrylic glass,

was reported to have a speed of sound of 2657 m/s,

attenuation of 5.3 dB/cm MHz and density of 1180 kg/m3

(Clarke et al. 1994; Hodgskinson et al. 1996; Tatarinov

et al. 2005). Epoxies and rigid polymers and polymer

composites can sufficiently approximate the acoustic

properties of bone. However, rigid polymers and

polymer composites are simpler models, whereas epoxy-

based materials can be more closely tailored to the desired

properties and configurations.

Trabecular boneTrabecular bone, residing within cortical bone, has

a porous structure that supports vascular tissues and

contains marrow. Trabecular bone is difficult to model

because of its tortuous framework and heterogeneity. In

most cases, trabecular bone phantoms have been designed

to contain bone marrow, and therefore the acoustic prop-

erties were tailored to more closely mimic marrow than

trabecular bone itself.

f hard tissue substitutes

tenuationdB/cm)

Impedence(MRayl) Source

– – Tatarinov 1998– – Tatarinov 1998– – Tatarinov 1998

.8 @ 1 MHz 8.4 Clarke et al. 1994; Tatarinov 19981 MHz 3.1 Clarke et al. 1994

@ 0.5 MHz – Clarke et al. 1994– – Barkmann et al. 2000

19 MHz 6.9 Singh et al. 2008– 17.0 Blodgett 2003

9 MHz 13.0 Singh et al. 2008

Fig. 3. Tubular specimens of ebonite, acrylic plastic, fiberglass and carbon fiber plastic (left) and layered cortical bonesubstitutes with rubber particles mixed in epoxy (right) (Tatarinov et al. 2005).


One group developed a trabecular bone substitute by

adding 1 mm cubic gelatin granules to liquid epoxy, de-

gassing and subsequently hardening the mixture (Clarke

et al. 1994). A range of porosities was achieved by varying

the volume of epoxy and gelatin, resulting in a speed of

sound range between 1844 and 3118 m/s and attenuation

between 7 and 17 dB/cm at 0.5 MHz (Clarke et al. 1994).

A gelatin and water mixture was used as the marrow

mimic. In another study, sunflower oil was embedded

into the pores of the material to act as a marrow substitute

(Strelitzki and Truscott 1998).

Trabecular bone phantoms have also been manufac-

tured by introducing holes into Perspex acrylic resins and

polyacetal materials, with the holes in the polyacetal filled

with water to mimic marrow (Hodgskinson et al. 1996;

Lee and Choi 2007). A phantom consisting of parallel

nylon wires, simulating trabeculae, was built in 2-D

rectangular grid arrays, with the thickness of nylon

wires chosen to match the trabecular thickness (Wear

2005). Nylon wires were previously shown to exhibit

frequency-dependent scattering similar to that exhibited

by trabecular bone (Wear 2004).

Whole boneWhole-bone phantoms include both cortical and

trabecular bone substitutes. A phantom composed of glass

beads dispersed in vulcanized silicone was used to assess

a tool for measuring mineral density in women (Young

et al. 1993). Polyvinyl chloride (PVC) tubes of varying

diameters and a speed of sound of 2300 m/s were used

as whole-bone phantoms in a study that used ultrasound

to gauge fracture risk (Barkmann et al. 2000). Axisym-

metric and nonaxisymmetric whole-bone phantoms were

made to assess cortical bone thickness using ultrasound-

guided waves, with axisymmetric phantoms built from

acrylic tubes filled with water (Moilanen et al. 2007)

and nonaxisymmetric phantoms made of PVC and filled

with butter (Moilanen et al. 2004, 2007). A fetal skull

bone phantom was built to validate the use of pulsed

Doppler ultrasound in studying cerebral vasculature and

was made from a high-density polyethylene (Vella et al.

2003). The phantom was reported to closely mimic the

fetal skull bone both acoustically and thermally (Pay

et al. 1998).

Finally, trabecular material was created using the

epoxy and sunflower oil technique, and a whole-bone

phantom was created by encasing it in a hollow Perspex

cylinder and degassing it in a vacuum chamber for studies

of osteoporotic fracture risk (Strelitzki and Truscott 1998).

Dental hard tissuesTeeth are primarily composed of enamel, the dense

fibrous ceramic composite on the outer tooth surface and

dentin, the inner structural material of a tooth that is

formed from a mineralized collagenous matrix. Enamel

was simulated using aluminum and dentin was simulated

using copper in a study that used laser-based ultrasound to

examine dental structure (Blodgett 2003). Aluminum was

found to closely match enamel in compressional (6300 m/s)

and shear (3100 m/s) wave speed of sound, as well as

acoustic impedance (17.0 MRayl), but copper was a poor

substitute for dentin.

Another group explored various glasses, ceramics

and metals as tissue substitutes for enamel. Soda lime

glass was ultimately selected because of its low attenua-

tion (6 dB/cm at 19 MHz) and comparable compressional

speed of sound (5789 m/s) and acoustic impedance (13

MRayl) (Singh et al. 2008). Self-curing resin-based dental

composite was selected as a dentin substitute in the study

over dental cements, epoxies and plastics because of its

moldability and its comparable acoustic properties (c 5

3306 m/s, A 5 108 dB/cm at 19 MHz, Z 5 6.9 MRayl)

to dentin. The group was able to prepare tooth phantoms

by injecting the composite material into a mold, curing


it and attaching the resulting composite ‘‘dentin’’ block to

a diced glass ‘‘enamel’’ slab with composite cement.

Cracks were embedded into the composite block during

curing, and dental restorations, including silver-mercury

amalgam fillings and gold and porcelain crowns, were

also integrated into the phantoms (Culjat et al. 2005;

Singh et al. 2007). However, unlike teeth with complex

internal microstructures, the phantoms were prepared

from two monolithic sections.

DISCUSSION AND CONCLUSION

Many soft tissue-mimicking materials have been

described that have a compressional speed of sound,

density, attenuation and acoustic impedance within the

measured range of soft tissues (Tables 1 and 2). The

backscattering coefficient, nonlinearity parameter and

shear wave speed of sound (in the case of hard tissue

substitutes) have rarely been reported and therefore are

not included in the tables. Agarose-based materials have

been the most widely used and are very well characterized

in the literature. They have the advantage that they are

simple to prepare, can be tailored to vary their acoustic

properties and are able to support a uniform distribution

of scatterers. However, agarose-based tissue substitutes

are limited in size (typically ,5 cm in thickness) because

they must have high surface-to-volume ratios to properly

congeal. The longevity of agarose-based materials is

highly dependent on handling and storage.

Polyvinyl alcohol–based materials have a more

complex fabrication process, which requires multiple

freeze-thaw cycles. Like agarose-based materials, PVA

materials can be acoustically tailored and can also support

a uniform distribution of scatterers. However, PVA mate-

rials have good longevity and structural rigidity, they can

be shaped and they are therefore the most attractive choice

among the soft tissue substitutes when longevity and

stability are of interest. Of the remaining tissue substitutes

described here, each is limited either by its acoustic prop-

erties, by its stability or by its structure. Open cell foam–

based materials cannot readily be acoustically tailored, but

are unique in that localized pathologies can easily be

embedded within a phantom. Oil gel–based tissue substi-

tutes may have promise but have not been sufficiently

characterized. Most gelatin-based substitutes have been

limited by low Young’s modulus and longevity.

Soft tissue phantoms have been described that have

incorporated many of the soft tissue-mimicking materials

described before. The most common tissue-mimicking

materials used in the fabrication of solid organ phantoms

have been those based from agarose, gelatin and PVA,

with a mixture of water and glycerin as the most common

blood substitute. Processes using agar and PVA were

developed that enabled multiple layers of soft tissues to

be combined to mimic multiple structures (Madsen et al.

1982; Frayne et al. 1993; Reinertsen and Collins 2006).

However, the bulk of soft tissue phantoms have had

homogenous internal structures. Soft tissue phantom

research efforts have focused mostly on vascular and

breast tissues.

Hard tissue phantoms have been developed using

epoxies, plastics and ceramics, and have recently begun

to advance with the advent of new materials and fabrica-

tion techniques. However, limited research has been

applied to the study of hard tissue substitutes to date,

and therefore a sufficient range of acoustic properties

has not yet been achieved. The majority of hard tissue

substitutes are simple and have good longevity, but their

acoustic properties cannot easily be tailored. Like soft

tissue substitutes, most hard tissue substitutes are homo-

geneous, and therefore lack the fibrous microstructure

and corresponding asymmetry present in hard tissues.

Of the hard tissue-mimicking materials described to

date, epoxies have the most promise. Epoxies can be

molded into the desired shape, and multiple studies have

demonstrated that epoxies can be combined with other

materials to achieve a range of acoustic properties

(Table 3) (Clarke et al. 1994; Tatarinov 1998).

Although numerous tissue phantoms are now avail-

able commercially, customized tissue substitutes continue

to have a role, primarily in the medical ultrasound research

community. Additional soft and hard tissue substitutes

will continue to be developed by academic and industry

research groups primarily because of the low cost and

design flexibility afforded by customized tissue-

mimicking materials.

Acknowledgment—Partial funding for this work provided by the Teleme-dicine and Advanced Technology Research Center (TATRC)/Depart-ment of Defense under award numbers W81XWH-07-1-0672 andW81XWH-07-1-0668.

REFERENCES

Anthony L, Aaron F. Theoretical and experimental quantification ofcarotid plaque volume measurements made by three-dimensionalultrasound using test phantoms. Med Phys 2002;29:2319–2327.

Bamber JC, Bush NL. Freehand elasticity imaging using speckle decor-relation rate. Acoust Imag 1996;22:285–292.

Barkmann R, Lusse S, Stampa B, Sakata S, Heller M, Gluer CC. Assess-ment of the geometry of human finger phalanges using quantitativeultrasound in vivo. Osteoporos Int 2000;11:745–755.

Bharadvaj BK, Mabon RF, Giddens DP. Steady flow in a model of thehuman carotid bifurcation. Part I—flow visualization. J BiomechEng 1982;15:349–362.

Blodgett DW. Applications of laser-based ultrasonics to the characteriza-tion of the internal structure of teeth. J Acoust Soc Am 2003;114:542–549.

Boote EJ, Zagzebski JA. Performance tests of Doppler ultrasound equip-ment with a tissue and blood-mimicking phantom. J Ultrasound Med1988;7:137–147.

Boyce KE. Development of a prototype anthropomorphic ultrasoundphantom: 1992 CIVCO/SDMS Innovation in Ultrasound Award.J Diagn Med Sonog 1993;9:32–37.


Brewin MP, Pike LC, Rowland DE, Birch MJ. The acoustic properties,centered on 20 MHz, of an IEC agar-based tissue-mimicking materialand its temperature, frequency and age dependence. Ultrasound MedBiol 2008;34:1292–1306.

Browne JE, Ramnarine KV, Watson AJ, Hoskins PR. Assessment of theacoustic properties of common tissue-mimicking test phantoms.Ultrasound Med Biol 2003;29:1053–1060.

Burlew MM, Madsen EL, Zagzebski JA, Banjavic RA, Sum SW. A newultrasound tissue-equivalent material. Radiology 1980;134:517–520.

Bush NL, Hill CR. Gelatine-alginate complex gel: A new acousticallytissue-equivalent material. Ultrasound Med Biol 1983;9:479–484.

Clarke AJ, Evans JA, Truscott JG, Milner R, Smith MA. A phantom forquantitative ultrasound of trabecular bone. Phys Med Biol 1994;39:1677–1687.

Culjat MO, Singh RS, Brown ER, Neurgaonkar RR, Yoon DC,White SN. Ultrasound crack detection in a simulated human tooth.Dentomaxillofac Radiol 2005;34:80–85.

D’Souza WD, Madsen EL, Unal O, Vigen KK, Frank GR,Thomadsen BR. Tissue mimicking materials for a multi-imagingmodality prostate phantom. Med Phys 2001;28:688–700.

Dabrowski W, Dunmore-Buyze J, Cardinal HN, Fenster A. A real vesselphantom for flow imaging: 3-D Doppler ultrasound of steady flow.Ultrasound Med Biol 2001;27:135–141.

Dabrowski W, Dunmore-Buyze J, Rankin RN, Holdsworth DW,Fenster A. A real vessel phantom for imaging experimentation.Med Phys 1997;24:687–693.

Davies RP, Kew J. Tissue phantom for learning US-guided vascularpunctures. J Vasc Interv Radiol 2001;12:267–268.

Dong F, Madsen EL, MacDonald MC, Zagzebski JA. Nonlinearityparameter for tissue-mimicking materials. Ultrasound Med Biol1999;25:831–838.

Edmonds PD, Ross WC, Lee ER, Fessenden P. Spatial distributions ofheating by ultrasound transducers in clinical use, indicated in a tissueequivalent phantom IEEE 1985 Ultrasonics Symposium, 1985;908–912.

Eriksson R, Persson HW, Dymling SO, Lindstrom K. Evaluation ofDoppler ultrasound for blood perfusion measurements. UltrasoundMed Biol 1991;17:445–452.

Erskine RL, Ritchie JW. Quantitative measurement of fetal blood flowusing Doppler ultrasound. Br J Obstet Gynaecol 1985;92:600–604.

Ferrara KW, Zager BG, Sokil-Melgar JB, Silverman RH, Aslanidis IM.Estimation of blood velocity with high frequency ultrasound. IEEETrans Ultrason Ferroelectr Freq Control 1996;43:149–157.

Frayne R, Gowman LM, Rickey DW, Holdsworth DW, Picot PA,Drangova M, Chu KC, Caldwell CB, Fenster A, Rutt BK. A geomet-rically accurate vascular phantom for comparative studies of x-ray,ultrasound, and magnetic resonance vascular imaging: Constructionand geometrical verification. Med Phys 1993;20:415–425.

Fromageau J, Brusseau E, Vray D, Gimenez G, Delachartre P. Character-ization of PVA cryogel for intravascular ultrasound elasticityimaging. IEEE Trans Ultrason Ferroelectr Freq Control 2003;50:1318–1324.

Garra BS, Insana MF, Shawker TH, Russell MA. Quantitative estimationof liver attenuation and echogenicity—normal state versus diffuseliver—disease. Radiology 1987;162:61–67.

Ghorayeb SR, Bertoncini CA, Hinders MK. Ultrasonography indentistry. IEEE Trans Ultrason Ferroelectr Freq Control 2008;55:1256–1266.

Giacomini A. Ultrasonic velocity in ethanol-water mixtures. J AcoustSoc Am 1947;19:701–702.

Hall TJ, Bilgen M, Insana MF, Krouskop TA. Phantom materials forelastography. IEEE Trans Ultrason Ferroelectr Frequ Control 1997;44:1355–1365.

Hodgskinson R, Njeh CF, Whitehead MA, Langton CM. The non-linearrelationship between BUA and porosity in cancellous bone. PhysMed Biol 1996;41:2411–2420.

Hopkins RE, Bradley M. In-vitro visualization of biopsy needles withultrasound: A comparative study of standard and echogenic needlesusing an ultrasound phantom. Clin Radiol 2001;56:499–502.

Hoskins PR, Loupas T, McDicken WN. A comparison of the Dopplerspectra from human blood and artificial blood used in a flowphantom. Ultrasound Med Biol 1990;16:141–147.

Hynynen K, Clement GT, McDannold N, Vykhodtseva N, King R,White PJ, Vitek S, Jolesz FA. 500-element ultrasound phased arraysystem for noninvasive focal surgery of the brain: A preliminaryrabbit study with ex vivo human skulls. Magn Reson Med 2004;52:100–107.

International Commission on Radiation Units and Measurements. Tissuesubstitutes, phantoms, and computational modelling in medical ultra-sound. Bethesda, MD: Author; 1998.

Jansson T, Persson HW, Holmer N, Sahlstrand-Johnson P, Jannert M.Ultrasound doppler for improved diagnosis of disease in the para-nasal sinuses. 2005 IEEE Ultrasonics Symposium 2005;2:839–841.

John C. The corono-apically varying ultrasonic velocity in human harddental tissues. J Acoust Soc Am 2004;116:545–556.

Jonsson P, Sahlstrand-Johnson P, Holmer N-G, Persson HW, Jannert M,Jansson T. Feasibility of Measuring Acoustic Streaming forImproved Diagnosis of Rhinosinusitis. Ultrasound Med Biol 2008;34:228–238.

Kawase Y, Suzuki Y, Ikeno F, Yoneyama R, Hoshino K, Ly HQ,Lau GT, Hayase M, Yeung AC, Hajjar RJ, Jang I- K. Comparisonof nonuniform rotational distortion between mechanical IVUS andOCT using a phantom model. Ultrasound Med Biol 2007;33:67–73.

Kerber CW, Heilman CB. Flow dynamics in the human carotid artery: I.Preliminary observations using a transparent elastic model. AJNRAm J Neuroradiol 1992;13:173–180.

Kharine A, Manohar S, Seeton R, Kolkman RG, Bolt RA,Steenbergen W, de Mul FF. Poly(vinyl alcohol) gels for use as tissuephantoms in photoacoustic mammography. Phys Med Biol 2003;48:357–370.

Kim YT, Kim HC, Inada-Kim M, Jung SS, Yun YH, Jho MJ,Sandstrom K. Evaluation of tissue mimicking quality of tofu forbiomedical ultrasound. Ultrasound Med Biol 2009;35:472–481.

Kondo T, Kitatuji M, Kanda H. New tissue mimicking materials for ultra-sound phantoms. Ultrasonics Symposium. 2005 IEEE 2005;3:1664–1667.

Landry A, Fenster A. Theoretical and experimental quantification ofcarotid plaque volume measurements made by three-dimensionalultrasound using test phantoms. Med Phys 2002;29:2319–2327.

Lee KI, Choi MJ. Phase velocity and normalized broadband ultrasonicattenuation in polyacetal cuboid bone-mimicking phantoms. J AcoustSoc Am 2007;121:EL263–EL269.

Lerski RA, Duggan TC, Christie J. A simple tissue-like ultrasoundphantom material. Br J Radiol 1982;55:156–157.

Liu J, Roberts CJ. Feasibility studies of model and system for ultrasoniccharacterization of cornea biomechanics. Invest Ophthalmol Vis Sci2004;45:U317.

Madsen EL, Berg WA, Mendelson EB, Frank GR. Anthropomorphicbreast phantoms for qualification of investigators for ACRINProtocol 6666. Radiology 2006a;239:869–874.

Madsen EL, Frank GR, Dong F. Liquid or solid ultrasonically tissue-mimicking materials with very low scatter. Ultrasound Med Biol1998;24:535–542.

Madsen EL, Frank GR, Krouskop TA, Varghese T, Kallel F, Ophir J.Tissue-mimicking oil-in-gelatin dispersions for use in heterogeneouselastography phantoms. Ultrason Imaging 2003;25:17–38.

Madsen EL, Hobson MA, Frank GR, Shi H, Jiang J, Hall TJ, Varghese T,Doyley MM, Weaver JB. Anthropomorphic breast phantoms fortesting elastography systems. Ultrasound Med Biol 2006b;32:857–874.

Madsen EL, Zagzebski JA, Banjavie RA, Jutila RE. Tissue mimickingmaterials for ultrasound phantoms. Med Phys 1978;5:391–394.

Madsen EL, Zagzebski JA, Frank GR. An anthropomorphic ultrasoundbreast phantom containing intermediate-sized scatterers. UltrasoundMed Biol 1982;8:381–392.

Madsen EL, Zagzebski JA, Ghilardi-Netto T. An anthropomorphic torsosection phantom for ultrasonic imaging. Med Phys 1980;7:43–50.

McDicken WN. A versatile test-object for the calibration of ultrasonicDoppler flow instruments. Ultrasound Med Biol 1986;12:245–249.

Meagher S, Poepping TL, Ramnarine KV, Black RA, Hoskins PR.Anatomical flow phantoms of the nonplanar carotid bifurcation,part II: experimental validation with Doppler ultrasound. UltrasoundMed Biol 2007;33:303–310.


Mo LY, Cobbold RS. A stochastic model of the backscattered Dopplerultrasound from blood. IEEE Trans Biomed Eng 1986;33:20–27.

Moehring MA, Ritcey JA. Sizing emboli in blood using pulse Dopplerultrasound. I. Verification of the EBR model. IEEE Trans BiomedEng 1996;43:572–580.

Moilanen P, Kilappa V, Nicholson PH, Timonen J, Cheng S. Thicknesssensitivity of ultrasound velocity in long bone phantoms. UltrasoundMed Biol 2004;30:1517–1521.

Moilanen P, Nicholson PH, Kilappa V, Cheng S, Timonen J. Assessmentof the cortical bone thickness using ultrasonic guided waves: model-ling and in vitro study. Ultrasound Med Biol 2007;33:254–262.

Nadkarni SK, Austin H, Mills G, Boughner D, Fenster A. A pulsatingcoronary vessel phantom for two- and three-dimensional intravas-cular ultrasound studies. Ultrasound Med Biol 2003;29:621–628.

Nicotra JJ, Gay SB, Wallace KK, McNulty BC, Dameron RD. Evalua-tion of a breast biopsy phantom for learning freehand ultrasound-guided biopsy of the liver. Acad Radiol 1994;1:385–387.

Oates CP. Towards an ideal blood analogue for Doppler ultrasound phan-toms. Phys Med Biol 1991;36:1433–1442.

Ophir J, inventor. Acoustic Standards Corporation, assignee. Ultrasoundphantom 1981. Patent number: 4,286,455.

Ophir J. Ultrasound phantom material. Br J Radiol 1984;57:1161.Patterson MS, Foster FS. The improvement and quantitative assessment

of B-mode images produced by an annular array/cone hybrid. Ultra-son Imaging 1983;5:195–213.

Pay NM, Shaw A, Bond AD. Evaluation of potential bone mimickingmaterials for ultrasound thermal test objects. Evaluation of potentialbone mimicking materials for ultrasound thermal test objects 1998;21.

Petrick J, Zomack M, Schlief R. An investigation of the relationshipbetween ultrasound echo enhancement and Doppler frequency shiftusing a pulsatile arterial flow phantom. Invest Radiol 1997;32:225–235.

Ramnarine KV, Anderson T, Hoskins PR. Construction and geometricstability of physiological flow rate wall-less stenosis phantoms.Ultrasound Med Biol 2001;27:245–250.

Ramnarine KV, Hoskins PR, Routh HF, Davidson F. Doppler back-scatter properties of a blood-mimicking fluid for Doppler perfor-mance assessment. Ultrasound Med Biol 1999;25:105–110.

Reinertsen I, Collins DL. A realistic phantom for brain-shift simulations.Med Phys 2006;33:3234–3240.

Rickey DW, Picot PA, Christopher DA, Fenster A. A wall-less vesselphantom for Doppler ultrasound studies. Ultrasound Med Biol1995;21:1163–1176.

Robbins CW, Kelly RM, inventors. Technicare Corporation, assignee.Biopsiable ultrasound phantom 1985. Patent number: 4,493,653.

Robertson J, Leen E, Goldberg JA, Angerson WJ, Sutherland GR,McArdle CS. Flow measurements using duplex doppler ultra-sound—hemodynamic-changes in patients with colorectal livermetastases. Clin Phys Physiol Meas 1992;13:299–310.

Robinson DE, Kossoff G. Performance tests of ultrasonic echoscopes formedical diagnosis. Radiology 1972;104:123–132.

Rowan M, Pedersen P. P2C-3 an injury mimicking ultrasound phantomas a training tool for diagnosis of internal trauma. Ultrasonics Sympo-sium, 2006 IEEE 2006;1612–1617.

Samavat H, Evans J. An ideal blood mimicking fluid for doppler ultra-sound phantoms. J Med Phys 2006;31:275–278.

Schaar JA, de Korte CL, Mastik F, van Damme LC, Krams R,Serruys PW, van der Steen AF. Three-dimensional palpography ofhuman coronary arteries. Ex vivo validation and in-patient evalua-tion. Herz 2005;30:125–133.

Sehgal CM, Bahn RC, Greenleaf JF. Measurement of the acousitcnonlinearity parameter B/A in human-tissues by a thermodynamicmethod. J Acoust Soc Am 1984;76:1023–1029.

Sheppard J, Duck FA. Ultrasonic tissue-equivalent materials using inor-gaanic gel mixtures. Br J Radiol 1982;55:667–669.

Shui G, Kim JY, Qu J, Wang YS, Jacobs LJ. A new technique formeasuring the acoustic nonlinearity of materials using Rayleighwaves. NDT&E Int 2008;41:326–329.

Singh RS, Culjat MO, Cho JC, Neurgaonkar RR, Yoon DC,Grundfest WS, Brown ER, White SN. Penetration of radiopaque

dental restorative materials using a novel ultrasound imaging system.Am J Dent 2007;20:221–226.

Singh RS, Culjat MO, Grundfest WS, Brown ER, White SN. Tissuemimicking materials for dental ultrasound. J Acoust Soc Am 2008;123:EL39–EL44.

Smith SW, Miller TM, Kisslo J. Anthropomorphic cardiac ultrasoundphantom with coronary arteries. Engineering in Medicine andBiology Society, 1991 Proceedings of the Annual InternationalConference of the IEEE 1991;13:138–139.

Smith SW, Rinaldi JE. Anthropomorphic cardiac ultrasound phantom.IEEE Trans Biomed Eng 1989;36:1055–1058.

Smith RF, Frayne R, Moreau M, Rutt BK, Fenster A, Holdsworth DW.Stenosed anthropomorphic vascular phantoms for digital subtractionangiography, magnetic resonance, and Doppler ultrasound investiga-tions. Medical Imaging 1994: Physics of Medical Imaging 1994;2163:235–242.

Sonotech. The Technology of Acoustic Scanning Gels. 2007.Strelitzki R, Clarke AJ, Truscott JG, Evans JA. Ultrasonic measurement:

An evaluation of three heel bone scanners compared with a bench-topsystem. Osteoporos Int 1996;6:471–479.

Strelitzki R, Truscott JG. An evaluation of the reproducibility and respon-siveness of four ‘state-of-the-art’ ultrasonic heel bone measurementsystems using phantoms. Osteoporos Int 1998;8:104–109.

Surry KJ, Austin HJ, Fenster A, Peters TM. Poly(vinyl alcohol) cryogelphantoms for use in ultrasound and MR imaging. Phys Med Biol2004;49:5529–5546.

Tatarinov A. Modeling the influence of mineral content and porosity onultrasound parameters in bone by using synthetic phantoms. MechCompos Mater 1998;35:147–154.

Tatarinov A, Sarvazyan N, Sarvazyan A. Use of multiple acoustic wavemodes for assessment of long bones: model study. Ultrasonics 2005;43:672–680.

Teirlinck CJ, Bezemer RA, Kollmann C, Lubbers J, Hoskins PR,Ramnarine KV, Fish P, Fredeldt KE, Schaarschmidt UG. Develop-ment of an example flow test object and comparison of five of thesetest objects, constructed in various laboratories. Ultrasonics 1998;36:653–660.

Tortoli P, Morganti T, Bambi G, Palombo C, Ramnarine KV. Noninva-sive simultaneous assessment of wall shear rate and wall distension incarotid arteries. Ultrasound Med Biol 2006;32:1661–1670.

Vella GJ, Humphrey VF, Duck FA, Barnett SB. Ultrasound-inducedheating in a foetal skull bone phantom and its dependence on beamwidth and perfusion. Ultrasound Med Biol 2003;29:779–788.

Wear KA. Measurement of dependence of backscatter coefficient fromcylinders on frequency and diameter using focused transducers—withapplications in trabecular bone. J Acoust Soc Am 2004;115:66–72.

Wear KA. The dependencies of phase velocity and dispersion on trabec-ular thickness and spacing in trabecular bone-mimicking phantoms.J Acoust Soc Am 2005;118:1186–1192.

Wells PN. Review: absorption and dispersion of ultrasound in biologicaltissue. Ultrasound Med Biol 1975;1:369–376.

Weskott HP. Amplitude Doppler US: slow blood flow detection testedwith a flow phantom. Radiology 1997;202:125–130.

Wojcik G, Szabo T, Mould J, Carcione L, Clougherty F. Nonlinear pulsecalculations and data in water and a tissue mimic. UltrasonicsSymposium, Proceedings IEEE 1999;2:1521–1526.

Xu D, Abbas S, Chan VW. Ultrasound phantom for hands-on practice.Reg Anesth Pain Med 2005;30:593–594.

Xu HX, Yin XY, Lu MD, Liu GJ, Xu ZF. Estimation of liver tumorvolume using a three-dimensional ultrasound volumetric system.Ultrasound Med Biol 2003;29:839–846.

Young H, Howey S, Purdie DW. Broadband ultrasound attenuationcompared with dual-energy X-ray absorptiometry in screening forpostmenopausal low bone density. Osteoporos Int 1993;3:160–164.

Zell K, Sperl JI, Vogel MW, Niessner R, Haisch C. Acoustical propertiesof selected tissue phantom materials for ultrasound imaging. PhysMed Biol 2007;52:N475–N484.

Zhang X, Greenleaf JF. Measurement of wave velocity in arterial wallswith ultrasound transducers. Ultrasound Med Biol 2006;32:1655–1660.

A Review of Tissue Substitutes for Ultrasound Imaging - McMaster

Documents